lkml.org 
[lkml]   [1999]   [May]   [31]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: XFS and journalling filesystems
On Sun, 30 May 1999, Steve Dodd wrote:
> I understand there's a difference between journalling and logging; guessing
> wildly, I'd say journalling is what you described, whereas logging is a
> system which stores incremental changes to filesystem state with 'last known
> good' checkpoints happening periodically. So in the latter case reading would
> indeed be slower as you'd have to look at all the changes that had occurred
> since the last checkpoint. Perhaps this is what was meant above.

I don't know if there IS an exact definition of journalling, but the way
I've come to know it, is that journalling is a big category of
filesystems, and log-structured fs'es are a sub-category. Normally,
journalling is done on meta-data, because (as you note) a fully
log-structured fs comes with a lot of design problems.

However, log-structured systems do use a garbage collector which cleans up
redundant data from the beginning of the log, and it can also defrag by
aggregating and moving blocks to the end of the log as it goes, releaving
the read penalties.

Still, all that fuzz for a HD-based filesystem is probably not worth it,
so that's why most journalled filesystems don't do data logging. Now,
there are other occations where this is desired, for example if the media
doesn't support random block writes (for example, standard flashes) - then
it's good to have a system that always writes sequentially.

/Bjorn



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:52    [W:0.121 / U:0.236 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site